Evaluating your SaaS onboarding process means finding the one step between signup and first value where most new users stop, and working out what that costs you in revenue. You need four things: a clear definition of an activated user, the journey to get there, a number for every step, and the reasons people drop off. This guide walks through it in seven steps you can run with your own data in a week.
What is a SaaS onboarding process?
A SaaS onboarding process is everything a new user goes through between signing up and reaching the first result your product promised them: the signup form, the first session, setup, guides, emails and any help from your team. It ends when the user gets value on their own, not when the product tour ends.
That distinction matters for the evaluation. A tour that 90% of users finish is not a success if they leave before they ever connect their data. You're evaluating the path to value, not the content along the way.
This is the user side of the picture. Our guide to what SaaS onboarding is covers the full definition, and customer onboarding covers what happens after a deal is signed.
How do you know if your onboarding process isn't working?
Usually the revenue tells you first. Signups are healthy, but trial-to-paid has been flat for months and no one can say why. Other signs:
- New users sign up and never come back for a second session.
- Your team keeps debating what "activated" means.
- Analytics is installed, but nobody trusts the events.
- Support gets the same setup questions every week.
- Customers cancel early, and you only find out in Stripe.
The cost is real. Userpilot's benchmark across 62 B2B SaaS companies puts the average activation rate at 37.5%: nearly two in three signups never reach first value. And early losses dominate: Focus Digital (2025) found that 43% of SMB customer losses happen in the first 90 days.
Which numbers should you check at each stage?
One number per stage is enough to start. Each comes from your product analytics or your billing data.
| Stage | The number to check | What it tells you | Where |
|---|---|---|---|
| Signup | Signup completion rate | Whether the form itself loses people | Analytics |
| First session | Share of signups who take the first key action | Whether new users know what to do next | Analytics |
| Setup | Share who finish setup (data connected, team invited) | Whether setup blocks the way to value | Analytics |
| First value | Activation rate and time to value | How many reach the result, and how fast | Analytics |
| Habit | Share of activated users still active in week 4 | Whether value turns into a routine | Analytics |
| PaidGoal | Trial-to-paid or free-to-paid conversion | Whether the routine turns into revenue | Billing |
Two benchmarks help you read them: the 37.5% average activation rate above, and a median time to value across SaaS of 1 day and 12 hours (Userpilot, 2025). Treat them as context, not targets. Your own trend, month on month, is the better guide.
How to evaluate your SaaS onboarding process in seven steps
1. Write down what "activated" means
Pick the action that best predicts a user who stays and pays: a first report run on real data, a second teammate invited, a first campaign sent. Test it against your own retention data. Users who took the action in their first week should retain far better than those who didn't. If your team can't agree on one action, that disagreement is your first finding. Our guide to activation metrics covers how to pick and track it.
2. Map the journey from signup to that action
List every step a new user takes, in order: the form, the welcome screen, the empty state, each setup task, every email they receive. Do it by signing up yourself with a fresh account. Then ask sales, support and customer success to walk it too. Each sees a different part. Sales knows the questions that come up in demos, support knows where people get stuck, and success knows which accounts never got going.
3. Measure every step
Turn the map into a funnel in your analytics tool, one step per event, and find the biggest drop. Then split it by acquisition channel, role and plan. An average hides a lot: an overall 25% activation rate can mean 45% for users from organic search and 8% for one paid campaign that attracts the wrong audience.
4. Find out why people stop
Numbers show where users drop off. They don't show why. For the biggest drop:
- Watch 10 to 20 session recordings of users who stopped at that step.
- Read the support tickets and chats from users in their first two weeks.
- Ask one question right after the step: "What almost stopped you here?"
- Call three users who signed up and never came back.
5. Compare against your past and the benchmarks
Look at the same funnel for each monthly signup cohort. A step that got worse after a release is easier to fix than one that was always weak. Then check the benchmarks above to see whether your activation rate and time to value are in a normal range.
6. Pick the one fix with the most revenue behind it
Rank the problems by what they cost: the users lost at that step, times your trial-to-paid rate, times your average contract value. Weigh that against the effort to fix it. Usually one step stands out. Fix that one first, not ten things at once, so you can tell what worked.
7. Ship it, measure against your baseline and repeat
Compare the cohorts that signed up after the fix with those before it. Small products rarely get enough traffic for a clean A/B test: across 127,000 experiments, Optimizely found that the average revenue-focused test lifts revenue by just 0.4%. A clear before-and-after on the step you changed is usually enough to decide. Then run the evaluation again next quarter.
What does a good result look like?
For Timelines.AI, the step that mattered was connecting WhatsApp, the moment trial users start getting value. We wrote onboarding emails aimed at that one step. In the 12 weeks after launch, the share of Zapier trial users who connected WhatsApp went from 47% to 55%. One step, one fix, one number to judge it by.
Your onboarding evaluation checklist
- Write down your definition of an activated user and test it against retention.
- Map every step from signup to that action, using a fresh account.
- Walk the journey with sales, support and customer success.
- Build the funnel and find the biggest drop.
- Split the funnel by channel, role and plan.
- Watch recordings and read tickets for the biggest drop.
- Compare this month's cohort with the last three.
- Put a revenue figure on each problem and pick one fix.
- Measure the fix against your baseline.
- Book the next evaluation for next quarter.
Should you evaluate onboarding yourself or get help?
Run it yourself if someone on your team has a free week and access to your analytics and billing data. The steps above are the whole method.
If no one has that week, or the data is messy, our 48-hour SaaS onboarding audit does the same evaluation on your product: your activation definition tested against your own retention, the funnel step by step, the leak in dollars and a ready-to-build spec for the top fix. If you'd rather talk it through first, book a 20-minute call with Jamie.
Questions
How often should you evaluate your SaaS onboarding process?
Check the main numbers monthly and run a full evaluation once a quarter, or straight after a big change such as a new pricing plan, a redesigned signup or a new kind of customer.
What is a good activation rate for a SaaS product?
Across 62 B2B SaaS companies, Userpilot found an average activation rate of 37.5%, so nearly two in three signups never reach first value. Your own trend matters more than the average. Compare each month's signups with the months before.
How long should SaaS onboarding take?
As long as it takes a new user to reach the first result they signed up for, and no longer. Userpilot puts the median time to value across SaaS at 1 day and 12 hours. If yours is measured in weeks, start there.
Do you need special tools to evaluate onboarding?
No. Product analytics (PostHog, Amplitude or Mixpanel), your billing data and a few session recordings cover most of it. The hard part is agreeing what "activated" means, not the tooling.
